Why SINAD is not important (article)

[JSmith]

SINAD is just as good as it gets

Nothing like a hot cup of SINAD in the morning;

JSmith

 

[TheTalbotHound]

One piece of bad measuring gear sounding bad to someone does not mean that they think good measurements imply bad sound. Indeed, Resolve seems to like a lot of the THX stuff, those measure well, and (i think) the Schiit Jotunheim 2, which measures pretty well too. So it seems that his issue is with the A90 specifically, and not simply with any gear that measures well.

 

[mfgillia]

Absolutely. This article does not present any new research at all, nor does it pretend to. It's simply a summary of findings across a large body of work. In that sense it should be treated more like a secondary or tertiary source, not some new grand and illuminating information on the usefulness (or lack thereof) of SINAD. The papers this article was based on are all referenced at the end, so any interested individual is free to read and interpret them.

Personally I think @Mad_Economist has defended his position well, and if anyone wishes to harp on the contents of the article at this point they should at the very least do a similar level of diligence. That is, refer to studies that support their case. I agree that the title comes across as sensationalistic, but it's intended for a layman's audience after all. A forgivable sin, if you will.

IMHO the title seems fine as it accurately summarizes the authors conclusions succinctly regardless whether one agrees with their conclusions about SINAD's relevance as an indicator for audio performance. Reading through some of the prior comments though, it appears that some equate questioning the value of SINAD on another reviewer's channel as an attack on measurements in general and ASR specifically.

However, there's nothing really new about this issue. Its been long debated on this site and pops up frequently. More recently this debate surfaced yet again during the Bluesound Node review, in which the product was ranked poor largely because of a SINAD below 96 even though most on both sides of that debate seemed to agree the low SINAD score was likely not audible.

 

[DonH56]

... I can assure you that low THD+N/SINAD in your master AD/DA signal chain will introduce unwanted, audible non-linearity elements, which are difficult to deal with later in the process.
<elided>

How, and what sort of non-linear elements?

Curious - Don

 

[MikeJ]

Let me elaborate a little bit more why I believe higher SINAD can tell you lot of things about the measured device in the first place. This will be in the context of DA converters. The higher the SINAD, the more I am confident that the victim DAC will have:
- very good noise floor / dynamic range (equals to competently designed input & output stages / power filtering)
- very low THD vs. output levels
- very low THD vs. frequency (can depend on reconstruction filter)
- very low IMD (i.e. the multitone test)
- very good linearity
- indication of very good clocking (i.e. jitter suppression)
Is this the case with poor SINAD measuring devices or just simply telling that SINAD is not important? I think you can answer for yourself.

 

[mdsimon2]

I think all of this is really good discussion.

I think most people do not understand how relatively inaudible distortion is (look at elevated 3rd harmonic of the SMSL M500 issue as a prime example).

I am somewhat troubled by the use of SINAD as support for dubious subjective listening results. I see a lot of people posting things like "wow, I went from this 110 dB SINAD DAC to this 120 dB SINAD DAC and it is so much clearer, darker blacks, smoother sound, etc". Ultimately this is more a data interpretation issue than anything else and overall ASR does a good job of interpreting the data in a realistic way. Although the SINAD chart is a bit silly if it wasn't provided by ASR someone would undoubtedly make their own based on the data.

Michael

 

[PierreV]

I see a lot of people posting things like "wow, I went from this 110 dB SINAD DAC to this 120 dB SINAD DAC and it is so much clearer, darker blacks, smoother sound, etc".

Indeed. One thing that bugs me a bit is when I see someone who uses a SOTA DAC (say SINAD of 120dB) on SOTA active speakers mentions

"a bit of hiss when I put my ear close to the tweeter, but inaudible from the listening position".

Obviously, without actual measurement of said hiss, we can't say by how much it raises the noise floor, but I always wonder by how much that hiss would reduce the SINAD anyway (assuming perfect speakers, which we all know don't exist).

It's probably a good thing that the speaker remains the limiting factor.

 

[MikeJ]

How, and what sort of non-linear elements?

Curious - Don

Yeah, it was a strange discovery. Once upon a time, if I recall correctly, it was circa when Lynx introduced its state of the art AD/DA converter to the audio world and impressed not only us with its linear AD conversion capability, just out of curiosity we have performed some loopback tests with several of our DA converters. For one of our "older" DA converters, after careful inspection of the recorded waveforms of test signals (with Lynx as the recorder), we have found out that the mentioned DAC was pushing a mix of strange odd and even harmonics to the signal itself. After more experiments, when the signal was re-recorded again, with only the DA/AD in the chain (to prevent coloration from other mastering gear), there was a distortion to the signal which, when compared to the results from the other DACs, sounded kind of strange to us (hard to describe, but certainly not musical at all). And you don't want this kind of coloration anywhere in your mastering chain in any part of the mastering process. We have first thought the AD is to blame, but that wasn't the case as the other converters from the test group didn't exhibit this behavior. Of course there were obvious level / noise differences between the resulting signals, but no such coloration as from this particular DAC was there. We have always thought that this specific DAC sounded somehow different (which wasn't really obvious when we ran the mix through several other mastering gear in the chain), but mix of summed, non-musical odd and even harmonics definitely shouldn't be there. That's it.

 

[mdsimon2]

Qa


Indeed. One thing that bugs me a bit is when I see someone who uses a SOTA DAC (say SINAD of 120dB) on SOTA active speakers mentions

"a bit of hiss when I put my ear close to the tweeter, but inaudible from the listening position".

Obviously, without actual measurement of said hiss, we can't say by how much it raises the noise floor, but I always wonder by how much that hiss would reduce the SINAD anyway (assuming perfect speakers, which we all know don't exist).

It's probably a good thing that the speaker remains the limiting factor.

100% agree with this. A bit funny that active speakers with noisy amps or ADC/DAC often get a pass but if those amps or ADC/DAC were measured directly they would be bottom of the barrel.

Michael

 

[KeithPhantom]

I'll give my thoughts later, but I can see that they are mixing the using of SINAD as a ranking of technical progress rather than auditory progress.

 

[DonH56]

Yeah, it was a strange discovery. Once upon a time, if I recall correctly, it was circa when Lynx introduced its state of the art AD/DA converter to the audio world and impressed not only us with its linear AD conversion capability, just out of curiosity we have performed some loopback tests with several of our DA converters. For one of our "older" DA converters, after careful inspection of the recorded waveforms of test signals (with Lynx as the recorder), we have found out that the mentioned DAC was pushing a mix of strange odd and even harmonics to the signal itself. After more experiments, when the signal was re-recorded again, with only the DA/AD in the chain (to prevent coloration from other mastering gear), there was a distortion to the signal which, when compared to the results from the other DACs, sounded kind of strange to us (hard to describe, but certainly not musical at all). And you don't want this kind of coloration anywhere in your mastering chain in any part of the mastering process. We have first thought the AD is to blame, but that wasn't the case as the other converters from the test group didn't exhibit this behavior. Of course there were obvious level / noise differences between the resulting signals, but no such coloration as from this particular DAC was there. We have always thought that this specific DAC sounded somehow different (which wasn't really obvious when we ran the mix through several other mastering gear in the chain), but mix of summed, non-musical odd and even harmonics definitely shouldn't be there. That's it.

Thanks! Sounds like a poor antialiasing and/or anti-imaging filter was letting stuff in/out that should not be there... Delta-sigma designs can exhibit strange-sounding signal/noise modulation so that is another possibility. Whatever it was, I doubt it was low distortion itself, but a consequence (probably unintended) of other design choices. But another case where low distortion on a single tone does not tell the whole story.

 

[MikeJ]

100% agree with this. A bit funny that active speakers with noisy amps or ADC/DAC often get a pass but if those amps or ADC/DAC were measured directly they would be bottom of the barrel.

Michael

I'm afraid there's little help for people out there who still believe that a "poor" SINAD DAC would be sufficient to power most of the current active speakers on the market with high fidelity. If you want transparency of your current active speakers, I would definitely start with the proper source, i.e. the DAC powering the active monitors itself.

 

[mdsimon2]

I'm afraid there's little help for people out there who still believe that a "poor" SINAD DAC would be sufficient to power most of the current active speakers on the market with high fidelity. If you want transparency of your current active speakers, I would definitely start with the proper source, i.e. the DAC powering the active monitors itself.

How does a “good” SINAD DAC help when you are limited by noise in the active speaker itself?

Of course if you have an active speaker with low noise / distortion internal electronics then a better DAC makes sense but that is not what we are talking about here. We are talking about cases where the active speaker has audible hiss (and hence internal electronics with poor SNR) being paired with a “good” SINAD DAC.

Michael

 

[MikeJ]

Indeed. One thing that bugs me a bit is when I see someone who uses a SOTA DAC (say SINAD of 120dB) on SOTA active speakers mentions

"a bit of hiss when I put my ear close to the tweeter, but inaudible from the listening position".

Obviously, without actual measurement of said hiss, we can't say by how much it raises the noise floor, but I always wonder by how much that hiss would reduce the SINAD anyway (assuming perfect speakers, which we all know don't exist).

It's probably a good thing that the speaker remains the limiting factor.

How does a “good” SINAD DAC help when you are limited by noise in the active speaker itself?

Of course if you have an active speaker with low noise / distortion internal electronics then a better DAC makes sense but that is not what we are talking about here. We are talking about cases where the active speaker has audible hiss (and hence internal electronics with poor SNR) being paired with a “good” SINAD DAC.

Michael

I don't think it's that simple to say it like that - we are talking about active monitors with amplifiers with their own THD+N contribution. Of course we are limited by acoustical properties of the speakers, but that's another discussion. Think about the active speaker amplifying "poor" / distorted DAC source signal vs. amplifying SOTA DAC signal. I'm not a speaker expert by any matter, but I'm 100% sure that the speaker membrane will "see" the final, amplified signal from the "poor" DAC as much worse signal (in "quality" terms), which will probably result in audible differences at certain listening levels. Noise is not imho the only limiting factor there - some distortion products can be buried inside noise, while the other ones will pass through anyway.

 

[mdsimon2]

I don't think it's that simple to say it like that - we are talking about active monitors with amplifiers with their own THD+N contribution. Of course we are limited by acoustical properties of the speakers, but that's another discussion. Think about the active speaker amplifying "poor" / distorted DAC source signal vs. amplifying SOTA DAC signal. I'm not a speaker expert by any matter, but I'm 100% sure that the speaker membrane will "see" the final, amplified signal from the "poor" DAC as much worse signal (in "quality" terms), which will probably result in audible differences at certain listening levels. Noise is not imho the only limiting factor there - some distortion products can be buried inside noise, while the other ones will pass through anyway.

I am not talking about acoustic properties of the speaker, as stated repeatedly I am talking about internal electronics of the active speaker.

With the level of hiss on some of these active speakers you will certainly be limited by the internal electronics. For example on the JBL 306 which has very audible hiss I’ve seen measurements of the internal amp (which has an integrated ADC/DAC) indicating SINAD of 55 dB at 5W. A high performance DAC upstream of that speaker isn’t doing anything for you.

Michael

 

[MikeJ]

Thanks! Sounds like a poor antialiasing and/or anti-imaging filter was letting stuff in/out that should not be there... Delta-sigma designs can exhibit strange-sounding signal/noise modulation so that is another possibility. Whatever it was, I doubt it was low distortion itself, but a consequence (probably unintended) of other design choices. But another case where low distortion on a single tone does not tell the whole story.

Could be, never bothered to use it anymore after the discovery. We have used single tones and full mixdowns for the test and looked at the differences. That strange distortion occurred just after two loopbacks. Don't have the DAC anymore but I'm sure when measured now, that mix of odd + even harmonics would definitely contribute to lower SINAD, as obviously there was something wrong within the DAC.

 

[JJB70]

There has often been something of a double standard in chasing SINAD in amplifiers and DACs (and deriding equipment with less than stellar SINAD) at the same time as ignoring audible hiss in active speakers or rationalising excuses for it. I actually agree that in almost all cases the hiss doesn't affect listening enjoyment and is outweighed by the benefits of a single well integrated package and I don't go around putting my ears against speakers looking for hiss anyway. But I think people should be honest about what it says if they suddenly see no issue with audible hiss.

 

[PierreV]

I don't think it's that simple to say it like that - we are talking about active monitors with amplifiers with their own THD+N contribution. Of course we are limited by acoustical properties of the speakers, but that's another discussion.

Indeed. Which is why I said "assuming perfect speakers".

which will probably result in audible differences at certain listening levels. Noise is not imho the only limiting factor there - some distortion products can be buried inside noise, while the other ones will pass through anyway.

Again, I agree this is a always a good idea to use a good DAC, which is why I said "assuming a SOTA DAC".

I was referring to the noise side exclusively.

I guess the gist of my point was that, it is a bit contradictory to worry about the difference between a 110 dB and a 120dB SINAD DAC and not worry about hiss because it is "inaudible" at the listening position.

 

[MikeJ]

100% agree with this. A bit funny that active speakers with noisy amps or ADC/DAC often get a pass but if those amps or ADC/DAC were measured directly they would be bottom of the barrel.

Michael

Missed this post. With speakers, @amirm does measure speaker's resolution in terms of frequency response, distortion, just to name a few metrics. It tells a lot about the performance of the speakers. Do you really think @amirm would measure only internals of the active speakers (in terms of SINAD) to make meaningful conclusions about speakers performance? It just doesn't make sense and you're taking things out of context. SINAD in active speakers amps is only part of the big picture, but of course important.

Ok, enough for now, hopefully we will put this debate to an definitive end...

 

[HiFidFan]

Ok, enough for now, hopefully we will put this debate to an definitive end...

Your optimism is admirable. . .